vlfrxtools/vtrtlsdr meteor detection.
Scott Shoemaker
Since “initiation” into meteor detection I have answered a couple of private messages regarding Linux and the vlfrxtools package so I thought I would expand in a post and also share some information on web based visuals.
Hardware is a Dell 15z laptop which you will see later is overkill.
RTL-SDR R820T – Best performer of three units tested along with E4000 and FC0013.
Antenna (today) is a TERK TV5 indoor. In the attic and oriented N/S.
Frequency 61.249.
OS is Ubuntu 12.04
Receiving software, and the meat of this post, vlfrxtools with the vtrtlsdr bundled (default download).
vlfrxtools is a very reliable and feature packed software solution and with it carrying the name, vlf, I think goes un-noticed as a program outside the range of vlf reception.
The recent addition of vtrtlsdr in the package opens the frequency range of dongles with the already impressive features included. Rather than go into the plethora of options I would highly recommend viewing the online documentation. Not included in on-line documentation is vtping, a meteor ping counter with graphic output.
In the program vtping -? Does give documentation.
The addition of vtrtlsdr gives the user association now to all plotting, graphics, and monitoring functions for gnuplot or already provided vtsid/vtsidex. Both sid and sidex are originally designed for VLF sid monitoring but with configuration modification fit meteor output data..
Be it known, this is not a GUI program and requires some scripting knowledge but if you’re like me, copy/paste works wonders and there are good examples in documentation or sharing. I am no Linux guru by any means but your favorite search engine can find solutions
Known challenges:
vtrtlsdr relies on librtlsdr and there are multiple posts that exist regarding buffer problems from the dongle. This is not unique to vlfrxtools and is widespread.
Most recommendations convey to push processor usage above 30%, and that solution does work on my Dell15z which is too much power for the task. If you want to use a lesser grade machine the vlfrxtools program is VERY lightweight and streamlined.
Future for my website – Time Doman plots, vtplot, join and plot time disciplined streams from my location with vtjoin.
If you already use vtrtlsdr with a NTP or better time disciplined PC, I would be open to ingesting one or two streams on my server with vtvorbis for stereo output, polar displays, and other comparisons.
Finally, I don’t want this to sound like I’m talking down on ARGO or Spectrum Lab (which I use in conjunction), PingOMeter, any of the other meteor tools. It’s just to convey a Linux program that from my experience works well and is one of the not so many meteor detection software programs.
Thanks, Scott - AD4TH
http://abelian.org/vlfrx-tools/
Alan Watson
--
You received this message because you are subscribed to the Google Groups "RadioMeteors" group.
To unsubscribe from this group and stop receiving emails from it, send an email to radiometeors...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Scott Shoemaker
vtrtlsdr -F 61.249e6 -r1200000 | vtresample -r11025 |
vtmix -c1,-j |
vtfilter -vvv -h bp,f=1000,w=600 -h lp,f=1200,poles=6 -h hp,f=700,poles=6 -g60 -- - @rtl0,30,i2
#vtrtlsdr- frequency 61.249Mhz dongle sample rate 1200000 and pipe
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; General Settings ;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; Specify directory to contain output files
; When running more than one copy of vtsid, each copy should have its own
; output directory.
datadir /raw/sid
; Specify the resolution. This can either be a number of bins, or a frequency
; in Hertz. Give one or the other. The number of bins doesn't have to be a
; power of two, but that is the most efficient. Otherwise, a number of bins
; which is the product of small factors is the next most efficient. But even
; a prime number of bins is handled reasonably well. If a resolution is given,
; this will be rounded to correspond to the nearest integer number of bins.
; resolution 1.0 ; Specify a resolution in Hz
bins 8192 ; Specify a number of frequency bins
; Specify how often monitor output records will be generated.
; Data will be averaged until the output interval is reached, then reported
; and cleared down.
; If this is set to zero, records will be generated as often as possible, which
; means every 1/resolution seconds, or every 2*bins/sample_rate seconds.
monitor_interval 5 ; Monitor output interval, seconds
; Indicate whether to monitor absolute phase. This requires the input stream
; to have a suitably accurate timestamp. Remove or comment-out if phase
; measurement is not required.
phase
; Spectrum records.
; Comment out the spectrum_interval statement to disable the spectrum records
spectrum_interval 120 ; Record the spectrum every 120 seconds.
; Specify the frequeny limits of the spectrum records. If the lower frequency
; is not given, the default is zero. If the upper frequency is not given,
; the default is the Nyquist frequency. The given frequencies will be rounded
; to a multiple of the resolution specified above.
;spectrum_lower 700 ; Lower frequency, Hz
;spectrum_upper 1300 ; Upper frequency, Hz
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; Channel Configuration ;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; Configure the input channels, indicating which channels are E-field or
; H-field and specify the effective azimuth of each H-field.
; Channels are specified in the order they appear in the input stream
; Format is
; channel type options
; Type is either efield or hfield.
; Options are a comma-separated list
;
; az= Specify effective azimuth of hfield source, mandatory for hfield.
; cal= Specify an amplitude calibration factor. The channel signal is
; multiplied by this factor before use. This can be used to balance
; a pair of loops, or to calibrate E/H ratio for elevation. The
; calibration factor defaults to 1.0
; H-field antennas need not be orthogonal, nor aligned with North/South or
; East/West.
;
; Examples:
; ; Single channel of E-field
; channel efield
; ; Pair of orthogonal loops, ch1 is N/S aligned, ch2 is E/W
; channel hfield az=0
; channel hfield az=90
; ; Non-orthogonal loops, not aligned with cardinal points.
; ; ch 1 aligned 35/215 deg, ch 2 aligned 145/325 degrees
; channel hfield az=35
; channel hfield az=145
; ; Orthogonal loops but not cardinal aligned, with E-field on ch3
; channel hfield az=45
; channel hfield az=135
; channel efield
channel efield cal=1.0
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; Monitor Settings ;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;
; The remainder of the config file is concerned with listing the signals to
; be monitored. Each monitor is identified by a case-sensitive 'ident' name.
;
; The format of each line is
;
; monitor ident type,options
;
; Type must be one of: noise, signal, msk, or cw
;
; The type determines how the channel is monitored and what parameters are
; to be measured and logged. Each type is described below.
;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; Noise bands:
; These are intended for measuring broadband noise and the average level of
; sferics. These monitors log only amplitude and bearing, and the amplitude
; is normalised to RMS per root Hz.
;
; Must specify center frequency and bandwidth using f= and w= options.
monitor Sferics7 noise,f=1000,w=600
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
; Signals:
; Ordinary signals which have no reliable phase for whatever reason. As with
; noise bands, only amplitude and bearing are logged, but the amplitude is the
; total RMS amplitude within the bandwidth.
;
; Must specify center frequency and bandwidth using f= and w= options.
; It is useful to monitor the hum levels
monitor hum50 signal,f=50,w=2 ; Monitor mains hum level
monitor hum150 signal,f=150,w=6 ; Mains 3rd harmonic
monitor MET signal,f=1000,w=50 ; Meteor SpectrumThe best instructions for graphical output is vtsidplot, vtsidgram, and vtsidex on the abelian site. I created a cron job to execute a .sh job with the parameters I wanted for every 15 minutes.
http://abelian.org/vlfrx-tools/notes.html#
You can also vtcat the source in timed (cron) intervals for multiple types of output, Just one example of many:
# Plot 30 S from the real-time data in @rtl0 vtcat -E0.30 @rtl0 | vtplot -t "Live Meteor" -o 'png large' > meteor.pngYou can substitute vtplot with any of the practical outputs, wspec, nspec, sgram.Streaming is straightforward:vtvorbis -ep -q0.4 -ktu shout,11.22.33.44,55,/meteor,xxxxx @rtl0Sends stream @rtl0 to server 11.22.33.44 port 55 mountpoint meter password xxxx-There are multiple suit to fit options in this description that may need to be tweaked to your setup.-For some reason some of the FC0013 sticks have swapped I/Q which requires modifying vtmix to get upper side band other wise you will have LSB.-All graphical output relies on gnuplot and subsequently can be modified with gnuplot arguments.-The vtvorbis to icecast already includes timestamping in the stream.-As with any un-modified dongle in any software there is thermal drift. Don't set your bandwidth to tight.http://www.ssbwx.net/VLF//sid/sidgram.pngDepending on your perspective, one negative here is timeliness. To be as close to live as possible I use Spectrum Lab and the capture option.The methods mentioned above obviously have to be first captured, then graphically produced. If you listen to the stream on my web pageyou hear ping and within seconds it is scrolling in the SL capture, pretty quick.As stated in the first post, I can host two maybe three Icecast connections on my server if anyone is interested, PM me.www.ssbwx.net:7777